Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 87
–B– The U.S. Vital Statistics System: A National Perspective National Center for Health Statistics INTRODUCTION The development and maintenance of a system to produce national vital statistics based on the local registration of vital events was a major accom- plishment of the United States during the 20th century. The National Center for Health Statistics (NCHS) is the federal agency legislatively mandated to produce national health statistics based on this cooperative, decentralized system in which data from more than 6 million vital-event records are col- lected each year by all states and U.S. territories and transmitted to NCHS for processing and dissemination. Looking to the 21st century, the local, state, and federal government organizations that comprise the National Vital Statistics System (NVSS) are engaged in significant changes through redesign and automation that should dramatically improve the performance and se- curity of the system. From data collection to data processing and dissemi- nation, efficiencies are being sought in an effort to improve both timeliness and quality at all levels of the system and to reduce costs. 87

OCR for page 87
88 VITAL STATISTICS THE NATIONAL VITAL STATISTICS SYSTEM In the United States, legal authority for the registration of births, deaths, marriages, divorces, fetal deaths, and induced terminations of pregnancy (abortions) resides individually with the states (as well as cities in the case of New York City and Washington, D.C.) and Puerto Rico, the Virgin Islands, Guam, American Samoa, and the Commonwealth of the Northern Mariana Islands. In effect, these 57 jurisdictions are the full legal proprietors of the records and the information contained therein and are responsible for main- taining registries according to jurisdiction law, including issuing copies of birth, marriage, divorce, and death certificates.1 As a result of this state2 authority, the collection of registration-based vital statistics at the national level has always depended on a cooperative re- lationship between the states and the federal government. Since its inception in 1960, NCHS has been the organization responsible for the federal aspects of this enterprise. NCHS has legislative authority and is mandated under 42 U.S.C. § 242k, Section 306(h) of the Public Health Service Act to collect vi- tal statistics annually: “There shall be an annual collection of data from the records of births, deaths, marriages, and divorces in registration areas. The data shall be obtained only from and restricted to such records of the States and municipalities which the Secretary, in his discretion, determines possess records affording satisfactory data in necessary detail and form.” Currently this data collection is limited to data from birth and death records (includ- ing fetal deaths), as NCHS discontinued the collection of individual-record marriage and divorce reports after 1995. The states are collectively represented in their dealings with the federal government by the National Association for Public Health Statistics and Information Systems (NAPHSIS). NAPHSIS is a professional organization whose members include primarily, but not exclusively, the vital registration and statistics executives and other employees of state registration offices. In addition to providing the states with a common point of contact with the 1 The NVSS is based on the local registration of vital events. For births and deaths, this typically works, in outline form, as follows: Demographic information on the birth certificate is provided by the mother at the time of birth, and medical and health information is based on medical (i.e., prenatal care, hospital, etc.) records. Demographic information on the death certificate is provided by the funeral director based on information supplied by the informant (usually the next of kin). A physician, medical examiner, or coroner provides medical informa- tion on cause of death. The completed birth and death certificates are registered with the local or state registrar by, respectively, the hospital records officer or the funeral director. The local registrar subsequently files the records with the state vital registration office, which codes and keys the data and transmits a copy of the electronic file to NCHS. The state offices are respon- sible for maintaining archival copies of records and for issuing certificate copies. Upon receipt at NCHS, the data are edited and assembled into national files for analysis and publication. NCHS sets uniform standards for data that will be collected and for item coding. 2 In the following text, the word “state” will be used to refer to all jurisdictions.

OCR for page 87
APPENDIX B 89 federal government and numerous other professional organizations, NAPH- SIS also facilitates interstate exchange of ideas, methods, and technology for the registration of vital events and dissemination of vital and other public health statistics. NAPHSIS’s progenitors date back to 1933, when it was or- ganized as the American Association of State Registration Executives; among other name changes, it was renamed the Association for Vital Records and Health Statistics (AVRHS) in 1980. MILESTONES IN NATIONAL VITAL STATISTICS The civil registration of births, marriages, and deaths has a long history in the United States, beginning with a registration law enacted by Virginia in 1632 and a modification of this law enacted by Massachusetts in 1639. The original impetus for these laws was the protection of individual rights, particularly relating to ownership and distribution of property, and not for statistical uses. However, with the rise of industrialism and urbanization in the 19th century, and the associated epidemics of diseases (typhus, yellow fever, cholera) arising from poor sanitation and polluted water, information from mortality records was used to develop support for sanitary reform and public health in general. The names of Edwin Chadwick (1800–1890) and Dr. William Farr (1807–1883) stand out in this effort in England, and that of Lemuel Shattuck (1793–1859) is prominent in the United States. Shat- tuck succeeded in steering through the Massachusetts legislature in 1844 a bill that required central state filing, provided for standard forms, fees, and penalties; specified types of information including causes of death; and lodged responsibility for each kind of record in designated officials (for more details, see Hetzel, 1997). In the late 1840s, the newly formed American Medical Association (AMA) began to promote the use of mortality statistics in the study of health conditions of the U.S. population. In 1847 the AMA addressed memorials to state legislatures on the need for registration laws. To obtain national data, the decennial censuses in the latter half of the 19th century included questions about vital events, but the method was recognized as inefficient and the results as deficient. Accordingly, in 1902, when the U.S. Bureau of the Census was made a permanent agency of the federal government, the legislation authorized the director of the Bureau to obtain, annually, copies of records filed in the vital statistics offices of those states and cities having adequate death registration systems and to publish data from these records. A few years earlier, the Bureau had issued a recommended death reporting form (the first “U.S. Standard Certificate of Death”) and requested each in- dependent registration area to adopt it as of January 1, 1900. Those areas that adopted the form and whose death registration was 90 percent com-

OCR for page 87
90 VITAL STATISTICS plete were to be included in a national death-registration area that had been established in 1880. In 1915 the national birth-registration area was estab- lished, and by 1933 all states were registering live births and deaths with acceptable event coverage and providing the required data to the Bureau for the production of national birth and death statistics. Fetal deaths of 20 weeks of gestation and greater have been a reportable component of U.S. vital statistics since the 1920s. Following the Supreme Court’s 1973 ruling on Roe v. Wade, the need for a separate reportable com- ponent on abortion was recognized and in 1978 the first standard report of induced termination of pregnancy (ITOP) was introduced. NCHS discontin- ued the collection of ITOP data from the states in 1994 as a cost saving mea- sure. Federally compiled data on abortions are available from the National Center for Chronic Disease Prevention and Health Promotion’s Division of Reproduction Health which has had a voluntary arrangement with states since 1969 to collect tabulated data on abortions, including the characteris- tics of abortion patients, occurring in the states. The current Model Law3 definition of fetal death specifically excludes ITOPs. There are more than 25,000 fetal deaths in the U.S. each year, similar to the number of infant deaths, and fetal death reporting remains an integral component of NVSS at NCHS. Registration areas for marriage (MRA) and divorce (DRA) statistics were established in 1957 and 1958, respectively. The MRA began with 30 states (excluding New York City) plus Alaska, Hawaii, Puerto Rico, and the Virgin Islands; when detailed marriage data were dropped in 1996 there was a to- tal of 45 registration areas. The DRA began with 16 states, the District of Columbia and the Virgin Islands; when detailed divorce data were dropped in 1995 there was a total of 33 registration areas. Beginning in 1996, NCHS has regularly published monthly counts of marriages and divorces in the reporting jurisdictions. Currently, all states are reporting counts of mar- riages to NCHS. Forty-four states and the District of Columbia are reporting counts of divorces (counts are not reported by California, Georgia, Hawaii, Indiana, Louisiana, and Minnesota). 3 A prime example of cooperative developmental work is the Model State Vital Statistics Act and Regulations, 1997 Revision; see http://www.cdc.gov/nchs/data/misc/mvsact77acc.pdf. NCHS facilitated the process of revising the Model Act and Regulations by convening a Re- vision Working Group and bringing in various experts that had an interest in the registration system. This is the fifth edition of the Model Act (the first was produced in 1907) and the third edition of the Model Regulations (the first was in 1973). The Model Act and Regulations provide a legal guide to states that are considering refining their laws. The report on the 1997 revision provides a useful reference on terminology, recommended registration practices, dis- closure and issuance procedures, and other functional aspects of a state registration system. The latest revision of the Model Act and Regulations contains provisions that allow states, which implement the relevant sections of the Model, to easily incorporate technological advances in records and information management.

OCR for page 87
APPENDIX B 91 In 1946, responsibility for collecting and publishing vital statistics at the federal level was transferred from the U.S. Bureau of the Census to the Fed- eral Security Administration, and later (1953) to the National Office of Vital Statistics (NOVS) in the U.S. Public Health Service. In 1960 NOVS was merged with the National Health Survey to establish NCHS. The Division of Vital Statistics (DVS) was created in a 1963 reorganization of NCHS. In 1987, NCHS became part of the Centers for Disease Control and Prevention (CDC) in the U.S. Department of Health and Human Services (HHS). THE VITAL STATISTICS COOPERATIVE PROGRAM Prior to 1971, the federal government reimbursed the states for use of their vital records for national statistics at the rate of 4 cents per record. The records were produced by local officials and transmitted to state vi- tal registration offices for permanent maintenance in central state files and production of state and local vital statistics. Basic standards for definitions, data quality, and methods of collecting, processing, and analyzing vital statis- tics were in place. The states forwarded microfilm copies of the records to NCHS, which edited, coded, tabulated, and published vital statistics for the nation. This process began to change in 1971, when the first state, Florida, transmitted magnetic tapes of state-coded data to NCHS, coded according to NCHS specifications. By 1973, six states had entered into contracts with NCHS to provide computer tapes of birth and demographic death data un- der a new formal arrangement called the Vital Statistics Cooperative Pro- gram (VSCP). During this early development period, priority was given to birth and demographic death data, and states were brought into the VSCP when they were technically ready and sufficient funds were available in NCHS to estab- lish a contract. Some states provided tapes to NCHS prior to availability of funding. Priority for contracts was given to non-VSCP states (rather than to adding subcomponents to states already under contract). The number of states submitting to NCHS one or more components of data under the VSCP gradually expanded as funding permitted. By 1985 all states, the District of Columbia, New York City, Puerto Rico, and the Virgin Islands were submitting birth data and demographic death (not cause of death) data on tape. In 1995, 42 states and the District of Columbia were submitting medical death data on tape. Except for periodic problems in reporting, all jurisdictions except West Virginia, Virgin Islands, American Samoa, Guam, and the Northern Marianas are currently submitting medi- cal death data in the format of the NCHS Mortality Medical Data System (described below).

OCR for page 87
92 VITAL STATISTICS The original VSCP contracts were negotiated individually with each state, but as states entered the VSCP it soon became apparent that a ratio- , nale had to be developed to simplify the process and assure funding equity among the states. In 1981 the director of NCHS established for this purpose a work group comprised of selected staff from NCHS and state representa- tives appointed by the president of the AVRHS. The work group’s assigned tasks were to (1) define the state/local activities involved in producing vital statistics for all levels of government, (2) figure the cost of those activities, (3) establish a rationale for determining the federal share of that cost, and (4) develop a standardized funding formula for the VSCP contracts. The end result of this effort was “a cost formula limited to the accepted level of effort necessary to carry out the in-scope functions of the vital statistics contracts.” This formula was used to distribute available funds to all states under VSCP contracts until 1987. A second VSCP work group was established in 1986 to review the ex- perience with the existing rationale and cost formula and develop recom- mendations for revisions. This work group recommended (1) updating and simplifying the cost formula to focus on the collection of standard data sets rather than an item-specific approach, (2) eliminating reference to the “fed- eral share” in favor of a funding level derived from a base level with annual cost-of-living adjustments, (3) providing for funding additions to cover the cost of collecting and processing new items of data, (4) providing for re- duction in scope of contract to offset reductions in federal funding, and (5) requiring a state to report all minimum basic data-set items to receive full contract funding. The third joint NCHS/AVRHS work group to review the VSCP cost for- mula was established in 1992. This work group made few changes in the pre- viously established VSCP provisions, recommending principally that the cost formula for the years 1995–1999 use the staffing and salaries data collected from the states in 1995 with overall application of annual cost-of-living ad- justments. This work group had extensive discussions of ways to improve timeliness of data production, ultimately recommending that states should send data to NCHS as soon as records were received and initially processed rather than waiting for full quality control to be completed; updated records were to be transmitted as amendments were processed. An important outcome of the emphasis on timeliness by the 1992 work groups was NCHS’s introduction of a new statistical series, based on a new approach to collecting and processing vital statistics data. Beginning with 1995 data, NCHS instituted an annual publication of preliminary vital statis- tics data based on a very substantial sample (80–90 percent) of records, including detailed tabulations from the natality as well as mortality files. Consequently, in January 1998, NCHS ceased publication of provisional mortality data based on the Current Mortality Sample (CMS), a sample of

OCR for page 87
APPENDIX B 93 10 percent of death records received each month from the states and coded and classified by federal staff. The Census Bureau had instituted the CMS in 1943 in response to concerns about the threat of epidemics and the pos- sibility of a general decline in national health resulting from wartime living conditions. In the year 2000, a new 5-year VSCP contract was established and sim- plified the cost formula that had been in effect since 1995, and this contract was extended in the years since 2006. For 2007 and 2008, CDC changed the procurement mechanism to purchase orders, but retained the funding distribution determined by the cost formula of 1995. Thus, the funding distribution to states remains basically the same as in 1995, with the total funding for the VSCP increasing each year by a general cost-of-living factor. Over the last 13 years, many changes in systems and procedures have taken place at the data provider and state levels, and it is unknown whether the current payment to states through the VSCP reflects an increase or decrease in the NCHS share of the current state cost of data collection and processing. The recommendations of the 1986 VSCP work group were applied in the mid-1990s as a result of funding reductions for NCHS. To adjust total contract funding downward, NCHS eliminated several VSCP components, ceasing collection of all data on induced terminations of pregnancy as of 1995 and detailed data on marriages and divorces in 1996. Several data items were also eliminated from the minimum basic data sets for natality (1-minute Apgar score, date of last live birth and last fetal death, and ed- ucation of father). For the mortality minimum data set, the autopsy item was removed, although later restored in 2002. As of 2008, however, other eliminated items and data sets have not been reinstated. Recent level-funding budgets for NCHS and other organizations in the federal government have created another crisis, but so far no data elements or components have been eliminated. Instead, the contracts (or purchase orders) have been funded to purchase data for shorter and shorter periods of time. Using the full VSCP cost for each year determined by successive cost-of-living adjustments, a daily “burn rate” is calculated for the fiscal year, and the contracts are then set up to cover the number of days in the year that the burn rate will purchase with the actual funding available. Complicating this picture is the issue of funding needed to cover the cost of important new medical data items in the 2003 revisions of the U.S. Stan- dard Certificates of Birth and Fetal Death. NCHS has held negotiations with NAPHSIS on this issue but, given NCHS’s inability to fund even a full year of the current minimum basic data set, there is little likelihood that funds will be available for additional data items in the upcoming budget years. This is a particularly unfortunate situation because the new medical and health data items are especially important for current issues in reproductive health. Currently, NCHS and NAPHSIS are engaged in the early stages of discus-

OCR for page 87
94 VITAL STATISTICS sions to develop strategies for disseminating data on these items, with focus on assessing data quality; the hope is that these collaborations will generate interest and support for steady and dependable funding. COOPERATIVE ACTIVITIES IN SYSTEM DEVELOPMENT E-Vital Initiative E-Vital was among the first of the 24 presidential e-government initiatives promulgated by the George W Bush administration and monitored continu- . ously by the U.S. Office of Management and Budget (OMB). The purpose of E-Vital was initially to establish common electronic processes for federal and state agencies to collect, process, analyze, verify, and share death and birth record information, thereby reducing the burden on state agencies for re- porting vital events and increasing the quality of the vital-event information being recorded. Until fiscal year 2008, the Social Security Administration (SSA) had the lead for this initiative and HHS, among other federal agen- cies, was a partner. During the initial phase of the initiative, a pilot system was developed to demonstrate the practicality of Electronic Verification of Vital Events (EVVE) by creating an online system that would enable federal agencies to query state vital records offices to verify the availability and accuracy of birth and death records, for example, to determine qualification for new or con- tinuing entitlements. The EVVE system was developed and shown to work; however, determining the appropriate payment per transaction for states re- mains elusive. Electronic Death Registration (EDR) The second portion of the E-Vital initiative, more important for vital statistics, was to support states in their efforts to reengineer their death regis- tration process, i.e., implement EDR. Timely reporting of death information is critical for detecting and defining pandemic and other calamitous events. Yet, timely death reporting has been a major challenge for the mortality vi- tal statistics system, primarily because state death registration systems have been essentially paper-based systems. Using grant funding (now exceeding $10 million) from SSA, states with assistance from NAPHSIS and NCHS have begun to implement EDR systems to improve the timeliness of fact-of- death information. Since 2002, 40 of the 57 registration jurisdictions (50 states, New York City, the District of Columbia, and 5 U.S. territories) have either imple- mented an EDR or are in the process of implementing one. The SSA ex- perience has successfully demonstrated that EDRs can provide fact-of-death

OCR for page 87
APPENDIX B 95 information from the local level through the state to the federal level in a timely fashion. However, state reporting of cause of death through these systems continues to be a challenge. The critical needs now are twofold: (1) expanding EDRs within and to all states, and (2) improving the timeliness of cause-of-death reporting through EDRs. Intelligence Reform and Terrorism Prevention Act In December 2004, the president signed into law the Intelligence Reform and Terrorism Prevention Act (IRTPA), which has the potential to boost ef- forts to implement electronic birth registration (EBR) systems as well as the EDR. Section 7211 of this act requires that the secretary of HHS establish minimum standards to improve the security of birth certificates. This section emanates directly from the 9/11 Commission’s report, which provided a va- riety of recommendations on terrorism prevention including the appropriate use and method of obtaining copies of birth certificates (National Commis- sion on Terrorist Attacks Upon the United States, 2004). Congress acted on this report with the passage of IRTPA, including Section 7211 which is the first federal statute to regulate vital registration practices of the states. IRTPA specifically mandates three categories of minimum standards for vital registration, including standards on (1) the certification of birth cer- tificates and the use of safety paper, (2) proof and verification of identity as a condition of issuance of a birth certificate, and (3) processing of birth certificate applications to prevent fraud. It also authorizes a grant program to assist states in meeting the federal standards and in computerizing their registration systems for the timely matching of birth and death records and noting the fact of death on the decedent’s birth certificate. While fund- ing was authorized for the states to implement the regulations, as yet no funds have been appropriated. It is anticipated that many states will need to modify some portion of their vital registration statutes in order to meet the proposed federal regulations. HHS was given the lead in the vital registration portion of IRTPA and the secretary of HHS delegated that lead to CDC; DVS within NCHS is handling the effort. All federal agencies affected by the legislation, including the U.S. Justice Department, the U.S. Department of Homeland Security, the U.S Department of State, the U.S. Department of Transportation, SSA, the Government Printing Office, and HHS, along with state vital registrars, were brought together to discuss the regulations. As a result of a series of meetings in the summer of 2005, this group has provided DVS with recommendations for these regulations. Using these recommendations and the services of a contract legal team paid for by SSA, DVS staff drafted a set of proposed regulations and submitted them to HHS for its review and approval.

OCR for page 87
96 VITAL STATISTICS The primary thrust of the draft regulations will be to standardize security practices for states in the registration and issuance of birth certificates. The goal is to have a more secure and responsive vital registration process in every state by enhancing their system’s infrastructure and providing secure electronic transmission of data within and between states and with federal agencies. The data will continue to be “owned” by the states; states will continue to manage and be responsible for vital registration; and there will be no federal database derived uniquely from this legislation. Despite the primary thrust on security, CDC/NCHS has a policy interest in these regulations and thus specific reasons to work in concert with other agencies to develop the regulations and adjudicate differences. Aside from the broader interest served by improving security, the primary policy inter- est of CDC/NCHS is to advance a long-standing public health interest in more rapid statistical information that is collected through the registration of births and deaths. Vital statistics have been built on the vital registration process in each state, and the modernization of the process and infrastruc- ture of vital statistics reporting (most specifically automating reporting at the source) can provide an early warning system in every community to track in real-time high-risk births and deaths of public health interest. COOPERATIVE ACTIVITIES IN DATA DEVELOPMENT U.S. Standard Certificates and Reports Periodic revision of the U.S. Standard Certificates and Reports is a sig- nificant area of cooperation for NCHS and the state vital statistics offices, occurring generally every 10 to 15 years. This is a particularly important ac- tivity because it brings together various experts—data users, researchers, and policy makers, both public and private—to develop recommendations on the content of the certificates and reports that will be used in the registration of live births, deaths, marriages, divorces, fetal deaths, and induced termina- tions of pregnancy during the next decade. Although the states do not all adopt the U.S. Standard Certificates and Reports exactly as they are promul- gated, the documents are generally employed with only minor changes and thereby succeed in promoting a high degree of uniformity and comparability among the states. In addition, VSCP contracts between NCHS and the states require collection of certain basic data items from the various certificates and reports. The standard certificates have been the principal means for achieving the uniformity in information on which national vital statistics are based. The U.S. Bureau of the Census developed the first standard certificates for the registration of vital events—births and deaths—in 1900. To date there have been 11 revisions of the Standard Certificate of Live Birth, 10 revisions

OCR for page 87
APPENDIX B 97 of the Standard Certificate of Death, 7 revisions of the Standard Report of Fetal Death (formerly Stillbirth), 3 revisions of the Standard Certificates of Marriage and of Divorce or Annulment, and 1 revision of the Standard Report of Induced Termination of Pregnancy. A published report (Tolson et al., 1991) describes the procedures followed in developing the revisions and the principal additions, modifications, and deletions of items for the 1989 revisions. This report also provides a history of the content of all certificates and reports since 1900. The 2003 Revisions The latest certificate revision process began in January 1998 and the re- port of the evaluation panel was issued in April 2000. Addenda to the report, explaining changes made after the initial recommendations of the panel, were issued in November 2001 (Division of Vital Statistics, 2000). The re- vision process evaluated only the live birth and death certificates and the fetal death report. The revised certificates and report were originally sched- uled for implementation in 2002. However, in consultation with NAPHSIS, NCHS decided to delay implementation until January 1, 2003, because of the complexity of changing automated systems in the states and the need to test the recommended changes before implementation. It was agreed that the fundamental goal should be to move from “a system primarily based on the flow of paper to the faster electronic registration of vital events. The Panel looked beyond designing new paper documents and concentrated on cultivating an appropriate vital statistics data base grounded in the electronic transfer of information” (Division of Vital Statistics, 2000). Moreover, it was recognized that the EBR systems in existence at the time were based on outmoded software and hardware and should be reengineered before new certificates were implemented. 2003 Changes in Collecting Data on Births and Fetal Deaths With the 2003 revision of the birth certificate and fetal death report, the panel recommended the development of worksheets to collect demographic data from the mother and separately to collect medical and health informa- tion from the prenatal care records and the birth facility. This was a signifi- cant shift from previous data collection procedures that typically depended on obtaining all relevant information from the mother. But this recommen- dation was subject to demonstration that data could be obtained accurately from the prenatal care records and the birth facility records by hospital staff. A study was successfully conducted in collaboration with NAPHSIS that found the information could be effectively collected this way. DVS also sponsored the development of a “Guide to Completing the Facility Work-

OCR for page 87
100 VITAL STATISTICS tion, with an LBIDF it is possible to use race of mother for both numerator and denominator in an infant mortality rate thereby improving the reporting of race for infant mortality data. Because an infant born in one state may die in another, the child’s birth and death certificates may be registered in different states. NCHS facilitates an interstate agreement to exchange infant death and birth certificates. LBIDF is produced in two different formats: birth cohort data and period data. The birth cohort data contain informa- tion for infants born in a particular year who died before their first birthday, either in that year or the following year. Beginning with 1995 data, NCHS also began producing period-linked file data. The period data include all infant deaths in a particular calendar year, linked to their respective birth certificates, whether the birth occurred in the current or the previous data year. The period format allows NCHS to release linked file data in a more timely fashion, since it is no longer necessary to wait for an additional year of mortality data to see if an infant died in the following data year. Thus, the period-linked file is a more effective tool for surveillance, while the birth- cohort-linked file is more suited to in-depth research projects. Birth cohort- linked file data are currently available for the 1983–1991 and 1995–2003 cohorts. Period-linked file data are currently available for 1995–2004 data years. Vital Statistics Follow-back Surveys The vital statistics follow-back surveys conducted by NCHS depend en- tirely on the cooperation of state vital statistics offices. A number of such surveys have been conducted since the mid-1950s by NCHS and its prede- cessor, the National Office of Vital Statistics, usually in collaboration with other federal government agencies. The state offices support the surveys by obtaining approval from their own health departments and institutional review boards, as necessary, and by authorizing the use of copies of vital records in the sample. These surveys typically are based on a sample of vital records in an annual birth or death file. Questionnaires are sent to sources of information identified on the records, for example, to the next-of-kin on the death certificate or to the mother on the birth certificate. The question- naires elicit additional information about the decedent or the mother and child, and in this way the survey provides a rich supplement to the infor- mation on the basic vital record. The sample data are weighted to provide unbiased estimates for the universe of records from which the sample was drawn. The 1988 National Maternal and Infant Health Survey (NMIHS) is the most recent follow-back survey conducted by NCHS focusing on repro- ductive health; it included a nationally representative sample of 9,953 live births, 5,332 infant deaths, and 3,309 late fetal deaths. This survey obtained

OCR for page 87
APPENDIX B 101 information on socioeconomic and demographic characteristics of mothers, prenatal care, pregnancy history, working history, health status of mother and infant, types and sources of medical care received, and lifestyle charac- teristics, including maternal smoking, drinking, and drug use. In 1991 NCHS conducted a Longitudinal Followup (LF) to the 1988 NMIHS by recontacting the mothers in the earlier survey to get informa- tion on their children’s health during the 2 or 3 years after birth. The LF requested data from the child’s medical providers as well. The LF provided information on infant feeding practices, child care, parental employment, and a wide range of information on early childhood health and develop- ment. A subsample of NMIHS women who had infant or fetal deaths (1,000 of each) was included in the LF survey to obtain data on subsequent repro- ductive behavior following infant or fetal loss. Since 1999, NCHS has been a collaborating agency with the Na- tional Center for Education Statistics on the Early Childhood Longitudinal Survey—Birth Cohort (ECLS-B). The ECLS-B examines children’s health, development, care, and education during the formative years from birth through kindergarten. Nearly 11,000 children, sampled from birth certifi- cates from across the United States, were included in the ECLS-B study. The latest National Mortality Followback Survey (NMFS) was conducted using death records for decedents who died in 1993. This survey included special samples of deaths from homicide, suicide, motor vehicle accidents, other accidents, HIV and certain natural causes. In addition to informa- , tion on the use of health services in the last year of life, the 1993 NMFS was unique because it included information obtained from the records of medical examiners and coroners. The NMFS also included information on socioeconomic and demographic characteristics of deceased persons, use of and payment for hospitals and institutional care during the last year of life, and various aspects of life style and other factors related to health status. In 2004, the Last Acts Partnership, a Washington, DC, advocacy organiza- tion for quality end-of-life care, recommended an ongoing NMFS (Last Acts Partnership, 2004), but to date no financial support has been forthcoming. The National Death Index The National Death Index (NDI) is a central computerized index of iden- tifying death-record information (beginning with 1979 deaths) at NCHS compiled from files submitted by state vital statistics offices. Working with state offices, NCHS established the NDI as a resource to aid epidemiologists and other health and medical investigators with their mortality ascertain- ment activities. Death records are added to the NDI file annually, approx- imately 12 months after the end of a particular calendar year. The NDI is available to investigators solely for statistical purposes in medical and health

OCR for page 87
102 VITAL STATISTICS research; it is not accessible to organizations or the general public for legal, administrative, or genealogy purposes. The NDI file contains a standard set of identifying information on each death to be used in searches of the file to identify and locate death records in the state offices. NDI users are encouraged to submit as many of the following data items as possible for each study subject: first and last name; middle initial; father’s surname; Social Security number; month, day, and year of birth; race; sex; marital status; state of residence; and state of birth. Results of NDI searches assist investigators in determining whether persons in their studies have died and, if so, provide the names of the states in which those deaths occurred, the dates of death, and the corresponding death cer- tificate numbers. Investigators can then either make arrangements with the appropriate state offices to obtain copies of death certificates or obtain cause- of-death codes using the NDI Plus service. Vital Statistics Training Program The NCHS vital statistics training program is another important activ- ity involving interaction with state personnel. DVS annually offers 1-week courses in “Vital Statistics: Measurement and Current Analytic Issues,” “Vi- tal Statistics: Measurement and Production,” and “Vital Statistics Records and Their Administration.” In addition, DVS staff has offered several types of courses on the coding and classification of cause-of-death information from death certificates. The participants in these courses are generally em- ployees of state, county, and city registration offices; however, mortality coders from other countries have also been trained in the use of the NCHS Mortality Medical Data System (MMDS). Over the past 20 years, more than 1,000 employees of these offices have taken one or more of these courses. NCHS ACTIVITIES IN DATA PRODUCTION AND DISSEMINATION Mortality Cause-of-Death Coding Cause-of-Death Classification Mortality statistics published as part of the NVSS are coded and clas- sified in accordance with World Health Organization (WHO) regulations, which specify that member nations use the current revision of the Interna- tional Classification of Diseases, ICD-10. ICD-10 not only details disease classification but also provides definitions, tabulation lists, the format of the cause-of-death section of the death certificate, and the rules for selecting the underlying cause of death. It provides the basic guidance used in virtually all countries for cause-of-death classification; the United States began using ICD-10 effective with deaths occurring in 1999.

OCR for page 87
APPENDIX B 103 WHO has provided a mechanism for updating the classification from time to time through the Update and Revision Committee and the Mortality Reference Group. Staff members in DVS are represented on both of these international groups and regularly attend group meetings, as well as the an- nual meetings of the heads of the WHO Mortality Classification Centers at various locations around the world. As a result of the periodic updates (mi- nor changes are made annually and major changes every 3 years), NCHS publishes updated versions of the ICD-10 tabular list. Mortality Medical Data System (MMDS) Beginning with the implementation of the eighth revision of the ICD in 1968, NCHS developed and employed several interrelated computer sys- tems to automatically select the underlying cause for each death certificate and to produce multiple cause-of-death data. System automation provides the benefit of greater consistency in the application of classification rules while requiring less extensive coder training. Currently, NCHS employs a suite of computer software, known as the MMDS, to code and classify cause- of-death information for most of the death records registered in the United States. The MMDS software is used by most states and many international partners to standardize the coding and classification of death records. There are three main software applications that comprise the MMDS. SuperMICAR accepts all literal entries of the certifying physician and au- tomatically converts the reported medical conditions into special numeric entity reference numbers (ERNs). The ERN output from SuperMICAR then becomes input to MICAR200 (Medical Indexing, Classification, and Retrieval), which assigns ICD codes to the ERNs for input to the third appli- cation, ACME (Automated Classification of Medical Entities), which in turn selects the underlying cause of death according to the rules of the ICD. (An- other program called TRANSAX uses the output from MICAR200 to pro- duce multiple cause-of-death data.) Also, SuperMICAR provides a method to retain literal entries in electronic form for quality control and analysis purposes of rare events subsumed under broad ICD codes. At present, the MMDS handles at least 85 percent of all death certificates; professionally trained classifiers code rejects from the MMDS manually. Not only utilized by NCHS and the states, MMDS is used in its totality by the following countries: Canada (English speaking), England and Wales, Scotland, Ireland, South Africa, and Australia. The following countries are completely dependent on ACME for determining underlying death cause: Sweden, France, Canada (French speaking), Hungary, and Brazil and will use other portions of the system in the near future. NCHS provides training and support with installation of the software as well as systems updating. Trinidad and Italy are in the process of implementing all or portions of the

OCR for page 87
104 VITAL STATISTICS system. Keeping these systems updated for the countries reflecting WHO annual changes has been a major support issue. NCHS also provides cause-of-death coding classes for our international partners. Over the last 5 years NCHS has trained staff from Isle of Mau- ritius, Kenya, Trinidad, Switzerland, Estonia, Italy, Poland, Spain, Latvia, Hungary, England, Czech Republic, Slovenia, Austria, and Tanzania. Matched Multiple Birth File DVS staff developed the Matched Multiple Birth File (MMBF) to facili- tate an analysis of characteristics of sets of births and fetal deaths in multiple deliveries. The MMBF currently includes six combined years (1995–2000) of data of matched sets of twins, triplets, and quadruplets in live births and fetal deaths. Live-birth records are linked to the corresponding infant death records for babies who died. Because of concerns for confidentiality with re- spect to small numbers for multiple births, some data fields are suppressed; no geographic identifiers are shown in the public-use version of this file. REPORTS AND PUBLICATIONS DVS statisticians and analysts produce a variety of publications and re- ports. There are standard reports that are produced annually from the na- tality and mortality data files that contain official statistics on U.S. births and deaths for a particular year. After the close of each calendar year, prelimi- nary birth and death files are produced. These are the basis for the reports “Births: Preliminary Data for Year XXXX” and “Deaths: Preliminary Data for Year XXXX.” Several months later, the files for the data year are closed and finalized. At that point, the reports “Births: Final Data for Year XXXX” and “Deaths: Final Data for Year XXXX” are produced. DVS analysts also produce annual reports on the linked birth/infant-death data set and the fetal and perinatal mortality data. The release of these “final” reports co- incides with the release of the public-use data files for these years. Annual reports are also produced on leading causes of death and life expectancy. Re- cently, staff members have introduced the public to the expanded health data from the 2003 revised birth certificate through an annual report, “Expanded Health Data from the New Birth Certificate.” DVS staff members respond to unexpected findings from the annual reports to produce special analyses. A recent example was in response to the unexpected increase in the infant mortality rate in 2002; the report was entitled “Explaining the 2001–02 In- fant Mortality Increase: Data from the Linked Birth/Infant Death Data Set.” DVS also produces a monthly report providing provisional counts of births, deaths, marriages, and divorces. DVS staff members regularly contribute to the CDC journal MMWR (Morbidity and Mortality Weekly Report).

OCR for page 87
APPENDIX B 105 In addition to these NCHS reports, DVS analysts have published in a variety of peer-reviewed journals. These articles may be coauthored with other NCHS analysts or with other CDC or federal colleagues or academic collaborators. The journals in which DVS staff have published in the last 2 years include Pediatrics, Paediatric and Perinatal Epidemiology, Interna- tional Journal of Health Services, Journal of Infectious Diseases, Seminars in Perinatology, Hispanic Journal of Behavioral Sciences, Journal of Marriage and Family, Fertility and Sterility, Birth: Issues in Perinatal Care, American Journal of Public Health, American Journal of Epidemiology, Demography, Maternal and Child Health Journal, Birth Defects Research: Clinical and Molecular Teratology, International Journal of Andrology, Diabetes Care, Di- abetologia, CA: A Cancer Journal for Clinicians, and Injury Prevention. DVS staff have also contributed invited chapters in a variety of books. DATA DISSEMINATION Releasing Microdata and Compressed Vital Statistics Files On November 20, 2007, DVS/NCHS released a new policy on the release of and access to vital statistics microdata for births, deaths, fetal deaths, linked birth/infant death, and matched multiple births. Effective with the 2005 data year, NCHS revised its microdata release and access policy to comply with state requirements, laws, and policies. This DVS revised policy reflects the dual goals to make data available as widely as possible while being responsive to concerns about confidentiality. Researchers can download public-use microdata files for births and linked birth/infant death data sets directly from the NCHS website, or they can be provided on CD-ROM or DVD.5 Birth, death, fetal death, and linked birth/infant-death public-use microdata files beginning with the 2005 data year contain individual-level vital-event data at the national level only, that is, with no geographic identifiers (no state, county, or city identifiers).6 5 For a complete statement of the DVS/NCHS microdata release policy, see http://www.cdc. gov/nchs/about/major/dvs/NCHS_DataRelease.htm. Microdata refers to records for individual cases. Files released by DVS may include a single record for each birth or death, or the file may be “compressed,” replacing identical records with a single record and the number of times that record occurs in the file. A compressed file reduces the number of records in the file. Compressed mortality files produced by NCHS list year and county of death, race (white, black, other), cause of death, and sex, use broad age groupings, and therefore do not contain as much detail as single-record microdata files. 6 Over the years, confidentiality standards have changed for the public release of geographic and date details on vital statistics microdata files. These changes are reflected in the data avail- able in successive time periods, as follows: • Birth, death, and fetal death public-use microdata files prior to 1989 contain all counties and exact dates (year, month, and day) of birth and death. • Birth, death and fetal death public-use microdata files for data years 1989–2004 con-

OCR for page 87
106 VITAL STATISTICS Researchers may request customized microdata files (birth, death, fetal death, and linked birth/infant death) and compressed files (death only) con- taining geographic detail for all states and counties for those data years with limited (1989–2004) or no (2005 forward) geographic detail in the public- use files. Data for approved projects are provided at no cost. Internet-Based Tabulation Query Systems Data users may also access data using Internet programs to construct their own tabulations of births and deaths with geographic detail subject to population or cell size limitations. Some of these interactive systems allow users to build tables based on microdata; however, only tabulated data are presented to the user. DVS has constructed an interactive tabulation system called VitalStats on the CDC/NCHS website. VitalStats is based on the Be- yond 20-20 software package. Users can tabulate, chart, and map natality, fetal death, and linked birth/infant-death data using prebuilt tables. They can also build their own tables based on natality and fetal-death data files. Trend tabulations of natality, fetal mortality, and linked birth/infant-death data by geographic detail at the county level are currently available.7 LOOKING AHEAD: VITAL STATISTICS FOR THE 21ST CENTURY Building on the Present Automation of Vital Statistics at the Source, State, and National Levels Despite the importance of the nation’s vital statistics system, in many states it remains based on outmoded vital registration practices and systems, tain only geographic identifiers of counties and cities with a population of 100,000 or greater, and no exact dates. For birth, death, and fetal death files, year, month, and day of week (e.g., Monday) are available. • Linked birth/infant-death public-use microdata files through 2004 contain geographic identifiers only for counties and cities with 250,000 or greater population and no ex- act dates. Year, month, and day of week (e.g., Monday) of birth/death are available. Beginning 2005 no geographic identifiers will be included on the public-use linked file. • The Matched Multiple Births File combines data from the 6 years 1995–2000 but ex- cludes all geographic identifiers and exact dates of births and deaths. The file also excludes year, month, and day of week (e.g., Monday). For a description of the file, see http://www.cdc.gov/nchs/r&d/rdc_twin.htm, and for download, ftp://ftp.cdc.gov/pub/ Health_Statistics/NCHS/datasets/mmb2/. An earlier version of this file, also available, combines data for 1995–1998. 7 Interactive systems currently available are VitalStats at http://www.cdc.gov/nchs/VitalStats. htm; WONDER (Wide-ranging ON-line Data for Epidemiological Research) at http://wonder. cdc.gov; WISQARS (Web-based Injury Statistics Query and Reporting System) at http://www. cdc.gov/ncipc/wisqars/; and IRHA (Interactive Reproductive Health Atlas) at http://www.cdc. gov/reproductivehealth/GISAtlas.

OCR for page 87
APPENDIX B 107 a fact that raises concerns about data quality, timeliness, and the lack of real-time linkage capabilities for the more than 6 million annual vital events. To resolve these issues, vital registration requires more complete automa- tion at the level of primary data collection and changes in the relationships among the providers of source records, the state registration offices, and NCHS. Collection of the cause-of-death information continues to be pri- marily a paper-based process, unchanged at the local and state levels for the last half century, and the reporting of fetal deaths is no better, if not worse. Although the states have been successful working with the funeral directors in automating the collection of demographic information on the decedent, there have been few successes in automation of the information collected from attending physicians, medical examiners, and coroners who provide and certify medical information on cause of death. The complete death-record data do not become computerized until reaching the state vital registration office, sometimes after considerable delay. The lack of automa- tion at the source precludes timely follow-back to improve data quality and does not take advantage of the expanded use of electronic medical records. Even the electronic sharing of information between states and with NCHS is rudimentary. To address these problems, NAPHSIS, NCHS, and SSA developed a part- nership to improve the responsiveness of state vital registration and statis- tics systems. Their objective was to improve the timeliness, quality, and sustainability of these systems by adopting national, consensus-based stan- dards and guidelines. Although these guidelines have been developed and have been used by some states to reengineer their birth registration systems and the demographic portion of their mortality systems, much remains to be done. Reengineered systems must include efficient methods for captur- ing data through standardized data collection instruments, coding specifica- tions, query guidelines, and definitions and transmit these data using HL-7 standardized messaging. As the Nationwide Health Information Network (NHIN) is knitted together, these reengineered vital statistics systems will need to be integrated with other electronic public health systems collecting information on immunizations, newborn screening, and hearing screening, and with electronic health records used by data providers, including hospi- tals and physicians. Many questions are yet to be answered. What is the most effective way to retrieve quality medical information from the attending physician, coro- ner, or medical examiner? How can funeral directors and physicians be connected electronically and share with the state confidential information about the decedent in a secure environment? At what level of specificity do prompts and data edits for the medical information obtained from the physi- cian become counterproductive? Efforts are currently under way to address some of these issues. SSA has been able to provide funding to some states

OCR for page 87
108 VITAL STATISTICS to automate the reporting of the fact of death, and NCHS is working to develop vital statistics data transmission standards. NAPHSIS is developing a data transfer system, yet the most daunting challenge remains the funding of the implementation of these new systems by the states. NCHS has had problems with funding the basic VSCP program and has been of little help in supporting states in their automation needs. As with the states, NCHS’s in- ternal systems also need to be completely reengineered to be able to interact with state systems on a real-time basis in order to follow back immediately to improve data quality as well as to publish and provide national vital statistics quickly on a year-to-date basis (Rothwell, 2004). Work is now under way to reengineer the internal systems within DVS to improve data quality and timeliness. Follow-back Surveys for Vital Statistics An Institute of Medicine and National Research Council (2003) report, Describing Death in America: What We Need to Know, “highlighted how little we know about ‘the quality, appropriateness or costs of care dying in- dividuals receive, or the burden on caregivers and survivors.” A reinstitution of the U.S. National Mortality Follow-back Survey (NMFS) could provide the information we must have if we are to improve care and plan intelli- gently for the future health care needs of our aging population. Such an on- going follow-back survey, taking advantage of new electronic health records and improved and linked vital statistics systems, could also on a regular ba- sis target causes of death of public health interest and more fully illustrate demographic disparities in mortality. Possibilities and Challenges The automation of vital statistics data collection at the source and its linkage to electronic health records and the building of responsive electronic transmission and linkage systems that will be required by the IRTPA legis- lation can provide significant new possibilities for how vital statistics are collected and utilized. For example, the 2003 revision of the birth record could be considered a rather extensive perinatal record. In the new environ- ment of automated and linked vital records and electronic health records, should all this information be collected for every record or should only a core of information be collected for each event augmented by a continu- ous follow-back survey or surveillance capability to sample electronic health records for information needed to elucidate certain health outcomes and health care practices? However, if birth certificates do not require this in- formation, will medical records contain the data in a standardized fashion, useful for sampling? With real-time access to mortality data and linkage to

OCR for page 87
APPENDIX B 109 electronic health records, is quicker annual reporting and quicker provision of annual mortality data files the primary outcome, or should this expanded capability be used to once again turn vital statistics into a dynamic public health surveillance system? To help with surveillance, could there be a pro- visional record collected followed by a more robust reporting of the causes of death for all or selected records? With the future dependence on elec- tronic health records and the growing need to depict chronic conditions, the use of the concept of the underlying cause of death may need to be revisited along with the automated medical coding systems, which turn literal medical nomenclature into an ICD code(s). The determination of appropriate contributions to a system that sup- ports a variety of government agencies is always difficult, and certainly this has been the case for the support of this nation’s vital statistics system. Com- plicating this situation is that although NCHS is authorized to collect vital statistics, states are not required to provide this information. If vital events are indeed vital should they not be required to be reported? Should a core data set be defined and required for each vital event and be made a re- portable event for states to provide to NCHS and then NCHS and its federal partners pay for follow-back on specific records of interest as well as training and maintenance of systems that support data sharing? Efforts to rejuvenate the nation’s vital statistics system are encouraging, but they will need to expand dramatically to provide a timely, high-quality, and flexible system to monitor vital health outcomes at the local, state, and national levels.

OCR for page 87